Success: kde 3.1 cvs running on sgi IRIX!

Thiago Macieira thiagom at
Mon Aug 12 03:06:46 BST 2002

Hash: SHA1

Tim Jansen wrote:
>On Sunday 11 August 2002 23:09, Waldo Bastian wrote:
>> Solution: don't use bzero but use the more portable memset instead. (Done)
>> Better solution: Remove this class and use the one from kdecore? (Tim?)
>Which? There is only KInetSocketAddress, which represents a host+port.

And in the case of IPv6, also the flowlabel and the scope ID.

>KInetAddress only represents the host address, and has the important ability
>to find out a address of an interface that connects to the local LAN and the
>address, that is used to communicate with the internet. That's why I wrote
>KInetAddress (and actually I planed to add it to kdecore in 3.2, because
>there are other apps that could use its functionality).

The address used to communicate can only be discovered with certainty when the 
communication is already established, via getsockname(2) and getpeername(2). 
And in the light of all the NAT that is being done, trying to discover one's 
IP address locally doesn't help much. As you said yourself in the comments, 
configuring the addresses using a KCM should be better.

As for the IPv6 addresses, they are only meaningful when connections are 
established. Not that computers don't have IPv6 addresses of their own -- 
quite to the contrary: they have lots. And that's exactly why trying to 
discover an address outside of a connection doesn't help much. (Just as a 
side note, my computer has 29 different global addresses).

I believe that the discovery of one's address has been discussed a while ago 
here, along with the gethostname(2) discussion.

BTW, I wrote KExtendedSocket to get rid of gethostbyname(3).

- -- 
  Thiago Macieira - UFOT Registry number: 1001
 thiagom at
   ICQ UIN: 1967141  PGP/GPG: 0x6EF45358
     Registered Linux user #65028
Version: GnuPG v1.0.7 (GNU/Linux)


More information about the kde-core-devel mailing list